# Zurich Instruments Newsletter - Edition Q1/2014

## Content

- Editorial
- Customer Interview: Toshiyuki Tsuchiya
- Reduce Lab Setup Complexity: The Periodic Waveform Analyzer
- Detailed Application Know-how: Detecting Vacuoles inside of Cells
- Premium Customer Care: Calibration Time
- Tips & Tricks: Phase Noise Measurements with a Lock-in Amplifier

## Editorial

In this edition we'll take a look at one of the options available with the 600 MHz UHFLI Lock-in Amplifier and also see how the UHFLI itself is being used in an interesting biological application - single-cell characterization. There is an interview with one of our HF2LI users in Japan, an explanation of how a lock-in amplifier can be used to make phase noise measurements and a reminder that now may be the time to get your instrument calibrated.

Finally there is a round up of the exhibitions and meeting we will be attending in the next few months - we look forward to meeting you around the globe!

## Customer Interview

### Toshiyuki Tsuchiya

**What is your most recent paper about?**

It is about the fatigue strength of sub-micrometer thick polysilicon thin films. We have developed a new vibration fatigue test system for membrane structures, which is free from the effect of patterning (DRIE, photolithography). This system is able to evaluate the intrinsic strength and fatigue properties of thin film materials. We observed a clear dependence on humidity of the fatigue lifetime.

**How many people have collaborated on this paper, and how was the collaboration performed?**

More than 6 people have worked on this paper. This is a joint research project with a company and the first author of the paper is from the company and also working with us as a PhD student. The company needed to evaluate the reliability of the polysilicon thin membrane structure, I proposed the method and the company and the student have developed a sophisticated test machine and specimen fabrication procedure. It was nice collaboration.

**Which experiment(s) are currently running (or setting up) and what are the objectives?**

We are now developing a MEMS optical chopper device. It is targeted to be used for time-resolved micro Raman spectroscopy to measure the dynamic local stress of vibrating silicon MEMS structures.

**Can you name one of your publications that has been referenced the most often?**

Tsuchiya T, Tabata O, Sakata J, Taga Y., Specimen size effect on tensile strength of surface-micromachined polycrystalline silicon thin films, J Microelectromech Syst 1998;7(1):106-13.

**How does the Kyoto environment influence your research?**

Kyoto is a historical city with many universities, colleges, and schools. This cultural environment encourages me to research more. And discussion with many excellent professors in Kyoto University inspires new ideas. I love Kyoto and our friends.

**At what age did you decide to become a physicist?**

Finally, at 18, when I entered university. I was interested in both physics and environmentology in my high school days. I liked to build electrical circuits and often went to Akihabara (Akihabara wikipedia) to buy electrical parts and made originally designed electrical equipment. Influenced by a teacher, I started to be interested in the conservation of nature. I wondered which way I should go, but finally I decided to become a physicist.

## Reduce Lab Setup Complexity

### The Periodic Waveform Analyzer

The Periodic Waveform Analyzer (PWA) is a type of synchronous detector which is useful for the analysis of periodic signals. Synchronous detection comes in a variety of different flavors and is a widespread concept often employed to optimize signal-to-noise ratio when extracting tiny signals from a noisy background. Different synchronous detection schemes all make use of the fact that the signal of interest is periodic and the associated frequency is a known reference. Lock-in detection and boxcar averaging are two practical examples of synchronous detection schemes, each tailored to a specific signal type - sinusoidal signals for lock-in detection and low duty cycle signals for boxcar averaging.

The PWA fits into this landscape by taking a continuous stream of input signal samples and associating each of these samples synchronously to the actual phase of the reference oscillator. Doing so over many periods of the reference frequency, leads to a very dense set of phase and sample pairs, where the phase is a value between 0 and 2π. This is a practical way of overcoming, by several orders of magnitude, the limits of the physical sampling rate of the ADC. Categorizing the phase values, for instance into 512 bins, leads to a waveform with sub-degree (i.e., 2π/512 or 0.7°) resolution, providing a phase domain representation. Moreover, a Fast Fourier Transform (FFT) operation applied to this dataset yields the frequency domain representation of the inputs signal at DC plus the first 255 harmonics with respect to the reference frequency.

Three typical use cases of the PWA are:

### Averaging of periodic signals

The reference frequency is either provided by the instrument itself or, when externally provided, carefully tracked with a phase-locked loop of adjustable bandwidth. Compared to standard oscilloscope signal analysis this offers the significant advantage of a trigger-jitter free averaging. Even more important is the possibility of dead-time free detection, whereas many real-time oscilloscopes have a considerable re-arm delay after each triggering event. In comparison, the Zurich Instruments UHFLI PWA can capture and average a data stream of 10 seconds or more, corresponding to 800 million periods for an 80 MHz laser without missing a single sample.

### Low duty-cycle signal analysis

The scope-like functionality mentioned above is also helpful to select a window of interest for boxcar-type measurements. In the phase domain representation one can easily pick the region that contains the valuable signal and, in case the resolution of 512 bins is insufficient, one can reference the input signal to a higher harmonic of the reference frequency which allows zooming into the region of interest and hence increasing the temporal resolution as much as is needed. That gives a fine analysis for pulsed signals with low duty cycles.

### Determination of efficiency of lock-in and boxcar measurements

The frequency-domain representation of the PWA, or to put it more accurately, the spectral distribution of the signal amplitude over the higher harmonics of the reference frequency, directly shows whether a lock-in or boxcar measurement is the best suitable detection method. Using the same reference oscillator as for the PWA and selecting the appropriate higher harmonic allows extraction of each single component of the spectrum displayed. Inversely, if one sees the signal spread out over many harmonic components without any prominent peak, a boxcar detection scheme might be the wiser choice to achieve best possible signal to noise ratio.

In summary, the Zurich Instruments UHFLI integrates a comprehensive toolset for synchronous detection. In particular, the PWA helps the user to exploit lock-in detection and boxcar averaging in the optimal way, all within one single commercial instrument.

## Detailed Application Know-how

N. Haandbæk, O. With, S. C.

Bürgel, F. Heer and A.

Hierlemann

### Detecting Vacuoles inside of Cells

The HF2IS Impedance Spectroscope from Zurich Instruments is a versatile tool for those doing label-free electrical impedance spectroscopy (EIS) for cell counting and cell discrimination. Today, the frequency range up to 50 MHz allows microfluidics researchers to not only to count but also detect the type of cells flowing through the microfluidic channel, with the multi-frequency capability making it possible to distinguish cell types and cell structures in a single pass.

More recently, several microfluidics labs have also employed the 600 MHz UHFLI Lock-in Amplifier. With more than a tenfold increase in the frequency range compared to the HF2IS, one microfluidic lab has raised EIS microfluidics to a new horizon with the help of UHFLI. As it can be seen in the plot below, measurement at tens of megahertz can at best characterize the impedance down to the level of cytoplasm inside a cell. However, at higher frequencies of several hundred megahertz, different strains of yeast can actually be distinguished based on assessing the impedance characteristics as a function of the number and the size of vacuoles [1].

In order to perform such vacuole-based discrimination of different yeast strains, the Department of Biosystems Science and Engineering from ETH in Basel (link) has devised two measurement setups which incorporate an UHFLI. In the first setup [2], the lock-in amplifier generates a multi-frequency signal to electrically excite the microfluidic chip electrodes. These signals are buffered by a pre-amplifier (PA). The PA is necessary since, at high frequencies, the mismatch between the signal generator output impedance (i.e. 50 ohm) and the microfluidic chip impedance will create significant reflections if not buffered. At the output electrodes of the microfluidic chip, a high frequency current-to-voltage converter (C2V) is also required in order to at the same time provide transimpedance gain and high frequency termination.

In the second setup [3], the microfluidic chip actually forms an LC resonance circuit with an external in-chip inductor. A change in impedance implies an immediate shift in the resonance frequency. Instead of measuring the impedance directly as in the previous setup, the impedance change can now be detected by tracking the resonance frequency shift using a phase-locked loop (UHF-PID Quad PID/PLL Controller). This setup targets improved measurement sensitivity due to the fact that the resonance frequency tracking requires less signal-to-noise ratio than the direct amplitude measurement.

[1] T. Sun and H. Morgan, "Single-cell microfluidic impedance cytometry: a review", Microfluid. Nanofluidics, vol. 8, no. 4, pp. 423-443, March 2010

[2] N. Haandbæk, S. C. Bürgel, F. Heer, and A. Hierlemann, "Characterization of subcellular morphology of single yeast cells using high frequency microfluidic impedance cytometer", Lab Chip, vol. 14, no. 2, pp. 369-377, January 2014

[3] N. Haandbæk, O. With, S. C. Bürgel, F. Heer and A. Hierlemann, "Microfluidic sensor using resonance frequency modulation for characterization of single cells", Proceedings of the 17th International Conference on Miniaturized Systems for Chemistry and Life Sciences, MicroTAS 2013, 2013, Freiburg, Germany, pp. 1680-1682

## Premium Customer Care

### Calibration Time

Did you know that Zurich Instruments offers an attractive calibration package for your HF2 or UHF Instrument?

Regular calibration of your scientific instrumentation provides several benefits for your daily measurement objectives including improved accuracy, reliability and confidence about the acquired data. Unfortunately parameter drifts of electronic component accumulate over time, reducing the accuracy of the instrument. Fortunately a re-calibration compensates for component aging and corrects the parameters back to the position where they are supposed to be. Measurements profit from optimal accuracy, certainly until the next calibration is due.

Calibrations also provide improved reliability, making the most of your valuable experimental time. How? Well, as the utilization of an instrument in a laboratory is often not tracked in detail, incorrect use or improper storage can lead to internal failures that might be difficult to detect. In the worst case scenario this can lead to false measurement results that an inexperienced user might not detect. Thanks to the servicing of the instrument that is part of a calibration, any use of your Zurich Instruments device can be performed with the confidence of the first day. Reliability is confirmed, thanks to the cleaning and visual inspection that's part of the servicing.

Your HF2 or UHF Instrument has a calibration interval sticker on the back panel of the device, so it is straightforward to determine when your measurement accuracy is potentially at risk. It is well understood that many experiments only require a precise instrument, but calibration is key for those who need optimal accuracy. That's why Zurich Instruments offers an attractive calibration and service package for its installed base, consisting of:

- Hardware inspection after several years of operation
- Calibration of long term drifts and renewed absolute accuracy measurements
- Additional warranty on electronic parts and workmanship, valid for 6 months after the recalibration
- Detailed calibration report
- Servicing of hardware, applicable hardware updates and removal of defects (additional charges may apply)
- Possibility to upgrade instrument with HF2LI-UHS or UHF-RUB options
- New remote training session, ideal for new research team members or as a "refresher"

Summarizing, a calibration service for your instrument has the potential to save valuable measurement time and improve the quality of laboratory life!

## Tips & Tricks

### Phase Noise Measurements with a Lock-in Amplifier

In the old days, phase noise measurement with a lock-in amplifier used to be performed by taking the FFT of the lock-in amplifier quadrature output Y. This is possible due to the fact that for very small phase variations, sin(Δϴ) ≈ Δϴ which can also be approximated by the variation in the measured Y amplitude. However, this approximation is not exact, especially if the phase variation is no longer 'very small'. Modern lock-in amplifiers like HF2LI or UHFLI directly output the measured phase by using the acquired in-phase component X and the quadrature component Y i.e. ϴ = arctan(Y/X). Therefore, the measured phase output can directly be taken to generate the phase noise profile of a measured signal without any approximations.

- In other words, the phase noise analysis is a plot of the phase noise power as a function of frequency. To calculate and plot the phase noise profile, one can use the procedure outlined below based on acquired data with the ziControl (HF2LI) or the LabOne (UHFLI) user interfaces.
- Save the measured phase value - the sampling rate Fs is set by the Readout Rate in the lock-in user interface
- Calculate the phase shift from the mean measured value i.e. Δϴ[n] = ϴmeas[n] – mean(ϴmeas)
- Determine the FFT length (nfft) to the nearest power of 2 and truncate Δϴ accordingly
- Apply a windowing function to contain spectral leakage in the calculation - for example, multiplying a Hann windowing function of nfft length with the truncated Δϴ data
- Perform the FFT with the windowed Δϴ samples to obtain the power spectral density PΔϴ - note that since FFT is symmetric, one only has to keep half of the FFT results, number of unique points nup = nfft / 2 + 1
- If the resulting data number nup is odd, then one needs to take PΔϴ from second point to the last point; if it is even, then take take PΔϴ from second point to the second last point. Remember to multiply PΔϴ by 2 since half of the spectrum energy was removed in the previous step
- Divide by the sampling rate mentioned in the first step to obtain spectral density information
- Determine the real frequency axis, Fnoise = (0 : nup - 1) * Fs / nfft
- Plot PΔϴ vs. Fnoise in 10log10 scale, the unit is dB rad2/Hz.

Please consider that the IEEE Standard 1139-1999 defines phase noise profile dBc as 3dB lower than the power spectral density PΔϴ.

The implementation is actually simpler than it reads. Feel FREE to contact [email protected] to obtain the related MATLAB code.